Validity of Using Background Leukemia Incidence Rates with Cohort Mortality-Based Potency Estimates to Calculate Excess Lifetime Risk
Data from occupational cohort mortality studies have been used to derive exposure-response curves and general population excess lifetime cancer risks, given low-level, chronic exposure. Using an actuarial method, mortality-based rate ratios associated with cumulative exposures are applied to age-specific background cancer mortality rates for a theoretical population aged birth to 70 years. In one recent U.S. Environmental Protection Agency health assessment, a mortality-based leukemia relative rate model was used with background leukemia incidence rates, rather than mortality rates, to calculate excess lifetime risk of leukemia incidence. We examined the validity and implications of this novel approach, while considering possible bias if a potential leukemogen did not pose equal risk by cell type. Limited sensitivity analyses were also conducted. Our analyses show that using total leukemia mortality-based potency estimates with background incidence rates will introduce a biased estimate of excess lifetime risk, the direction of which varies by potency and the histological type of leukemia. These biases were somewhat increased on adjustment for possible greater susceptibility of children. For potent carcinogens, the traditional approach provides a reasonable approximation of excess lifetime mortality risk for both the more and less fatal forms of leukemia, even after adjustments for children and is, therefore, to be preferred. Less consistency by leukemia cell type and background rate was observed for flatter exposure-response curves. This evaluation illustrates the importance of carefully examining the impact of methodological changes to calculations of excess lifetime risk before implementation.